Surrogate Models

From a mathematical point of view, simulations of chemical production processes involve the solving of large non-linear equations systems. The solution set – from the process view the possible operating window of a plant – is thereby a priori unknown, but hidden in the structure of the equation system. Here, AI methods shall be implemented in order to learn the solution set as a first step. It represents the operation window of the production process. With this knowledge, it will be possible to identify significantly better plant designs and operation strategies than previously done.

Contact: PD Dr. Michael Bortz, ITWM


Michael Bortz, Fraunhofer ITWM